--- a +++ b/inst/benchmark_stmm.m @@ -0,0 +1,78 @@ +% Copyright (C) 2008 Jaroslav Hajek <highegg@gmail.com> +% +% This file is part of OctaveForge. +% +% OctaveForge is free software; you can redistribute it and/or modify +% it under the terms of the GNU General Public License as published by +% the Free Software Foundation; either version 2 of the License, or +% (at your option) any later version. +% +% This program is distributed in the hope that it will be useful, +% but WITHOUT ANY WARRANTY; without even the implied warranty of +% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +% GNU General Public License for more details. +% +% You should have received a copy of the GNU General Public License +% along with this software; see the file COPYING. If not, see +% <http://www.gnu.org/licenses/>. +% + +% function benchmark_stmm (n, nvec) +% description: +% Sparse transposed matrix-vector multiplication benchmark. +% This is to test the "compound operators" feature introduced in Octave. +% +% arguments: +% n = dimension of matrix +% nvec = number of vector op repeats +% +% results: +% time_tmm = Time for A'*B (B n^2-by-nvec matrix) +% time_tmv = Time for A'*v nvec-times (v vector) +% time_mtm = Time for B*A' (B nvec-by-n^2 matrix) +% time_mtv = Time for v*A' nvec-times (v vector) +% + +function results = benchmark_stmm (n, nvec) + + benchutil_default_arg ('n', 300); + benchutil_default_arg ('nvec', 100); + + benchutil_initialize (mfilename) + + disp ('constructing sparse matrix') + n = 300; % size of the grid + m = n^2; % number of points + X = (n-1)*rand (m, 1); Y = (n-1)*rand (m, 1); + IX = ceil (X); JY = ceil (Y); + + A = sparse(m, n^2); + A = A + sparse (1:m, sub2ind ([n, n], IX , JY ), (IX+1-X).*(JY+1-Y), m, n^2); + A = A + sparse (1:m, sub2ind ([n, n], IX+1, JY ), (X - IX).*(JY+1-Y), m, n^2); + A = A + sparse (1:m, sub2ind ([n, n], IX , JY+1), (IX+1-X).*(Y - JY), m, n^2); + A = A + sparse (1:m, sub2ind ([n, n], IX+1, JY+1), (X - IX).*(Y - JY), m, n^2); + + v = ones (m, nvec); + tic; u = A'*v; time_tmm = toc; + benchutil_set_result ('time_tmm') + + v = ones (m, 1); + tic; + for i=1:nvec + u = A'*v; + end + time_tmv = toc; + benchutil_set_result ('time_tmv') + + v = ones (nvec, m); + tic; u = v*A'; time_mtm = toc; + benchutil_set_result ('time_mtm') + + v = ones (1, m); + tic; + for i=1:nvec + u = v*A'; + end + time_mtv = toc; + benchutil_set_result ('time_mtv') +